ABSTRACT ? Project 2 Developing an understanding of the effect of variability in vitamin K intake over time on the relationship between key genes and warfarin related outcomes is important for several reasons. First, the majority of pharmacogenomics research on warfarin over the past decade plus has focused on CYP2C9 and VKORC1 during treatment initiation. While these genes have been shown to influence INR, warfarin dose, and bleeding risk during the initiation period, a recent randomized clinical trial conducted in the US of a pharmacogenomic dosing algorithm found no significant improvement in INR during the first month of warfarin therapy. To ultimately provide information that enables clinicians to improve patient outcomes beyond that provided by current (already individualized) standard care, effects beyond those already studied need to be identified. Second, although patients are at highest risk of adverse bleeding events during the first few months of treatment, the duration of exposure to risk of both clotting and bleeding is significantly longer during the maintenance phase of treatment. Clinicians and patients need information that can help improve the safety of warfarin during the extended maintenance phase of therapy, during which variability in environmental exposures, rather than dose-finding, places them at risk of anticoagulation related events. To complement our previous evaluation of bleeding risk during the maintenance phase of therapy, in Aim 1 we ask the following question: If CYP4F2 is protective for major bleeding in long-term warfarin patients, is it a risk factor for venous thromboembolism (VTE)? We will employ an efficient case-control study design using electronic medical records to address this question. We will conduct in Aim 2 both genome-wide analyses and candidate gene analyses informed in part by work in Project 1, including the use of innovative machine learning algorithms. Identification of novel associations with VTE risk will be valuable for informing future research on both genetic and environmental effects on the vitamin K cycle, as well as potentially informing clinical care. Finally, in Aim 3 we have designed a prospective, longitudinal study of warfarin patients in the maintenance phase of therapy to collect data on quantitative vitamin K intake, genetic variation, PIVKA-II levels, and both warfarin dose and INR. Measurements will be taken 4 times over the period of one year to capture seasonal variation in vitamin K intake. Patients from three sites (Washington, Montana, Alaska) of AI/AN and of European ancestry will be recruited to capture geographic and cultural differences in vitamin K dietary intake, as well as genetic diversity. This study design will enable us to carefully evaluate the impact of dietary vitamin K on phenotypic markers and surrogate clinical outcomes, and will directly assess our hypothesis that variation in dietary vitamin K interacts with the influence of genetic variation on warfarin-related clinical outcomes.